Recognizing Cassava Variety Using Artificial Neural Network with Otsu Algorithm for Image Segmentation
Magdalene C. Unajan1, Bobby D. Gerardo2
1Magdalene C. Unajan, Graduate Programs, Technological Institute of the Philippines, Cubao, Quezon City, Philippines.
2Bobby D. Gerardo, College of Information and Communications Technology, West Visayas State University, Iloilo City, Philippines.
Manuscript received on 13 March 2019 | Revised Manuscript received on 20 March 2019 | Manuscript published on 30 July 2019 | PP: 131-135 | Volume-8 Issue-2, July 2019 | Retrieval Number: A1917058119/19©BEIESP | DOI: 10.35940/ijrte.A1917.078219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: One of the many indispensable tools for ensuring a quality product is variety identification. Human experts identify the product variety by personal observation; however, in their absence and rarity, technology can be used instead. This paper proposed to develop a technique that can be used to determine the type of cassava through its digital leaf image. There are 235 images used for testing and preprocessing. Two images for each of the 47 cassava varieties are used in this study. The preprocessing method was performed first before the extraction of features. The Otsu algorithm segments the leaf image from the background. From the leaf samples, nine (9) color features, three (3) morphological features and, three (3) shape features were extracted. The values of the 15 extracted features are the input for the system for variety recognition. Backpropagation method of the artificial neural network (ANN) of multilayer perceptron is used to train the system. For the input, hidden, and output layers, the values are 15, 30, and 47, respectively. These correspond to the 15 extracted features, 30 hidden layers, and 47 cassava varieties. The accuracy obtained in the experiment is 85.11%. It can be concluded that the technology was able to identify the different cassava varieties effectively.
Index Terms: Artificial Intelligence, Backpropagation, Feature Extraction, Image Segmentation, Precision Agriculture
Scope of the Article: Image Security